CHAPTER 1 CIRCUITS Section 1. Vacuum Tube Basics 1. Introduction During the 1940s, when the ENIAC was under construction, the basic principles of vacuum tubes and their circuits were familiar to most electrical engineers. Today, however, that knowledge is far less common. For that reason, we begin our discussion of the circuit design of the ENIAC with a brief look at how vacuum tubes work and how they are used as switching elements in circuits. Conversely, the reader is expected to be somewhat familiar with the functions of other circuit components, especially resistors and capacitors. 2. Tube Structure As the name implies a vacuum tube operates in a vacuum. This allows the electrons (e−) to flow freely between the electrodes in the tube without being impeded by gas molocules. That vacuum is usually contained in a glass envelope, though sometimes we find them in metal cans. These devices contain at least two electrodes, and those that are used in logic circuits that we study here all have at least three. a. Cathode. The cathode is the electrode from which the electrons origi- nate. To cause it to give off electrons, its temperature is raised with a heater. In some tubes the heater is the cathode, but for the tubes used in the logic of the ENIAC, the heater is inside a small metal can. This can is coated with a mixture of oxides, such as those of barium and strontium, that efficiently emit electrons at lower temperatures than the heater itself. b. Plate. The plate (or sometimes anode) is physically designed as a metal can surrounding the cathode with some free space between them. When the plate is at a sufficiently high voltage with respect to the cathode, then the electrons emitted from the cathode travel across the gap between them and hit the plate. When this happens, a current flows through the tube. Note, that the electrons can only flow from the cathode to the plate and not the other way around. A tube containing just a cathode and a plate is called a diode. c. Grid. Tubes can also have one or more grid electrodes between the cathode and the plate. If there is a single grid, then the tube is called a triode, if it has two grids, then it is called a tetrode, and if it has three grids, it is called a pentode. Physically, the grid is usually formed as a wire spiral around the cathode and in the space between it and the plate. The spiral is loose enough 1 PLATE GRID HEATER CATHODE Figure 1. Triode Schematic Symbol that there is space between the wires for the electrons to pass. If the grid is at the same potential as the cathode, then other than some electrons colliding with the grid wires, it is as if the grid is not there, and current can pass. If the grid is at a negative potential with respect to the grid, then the negative charge around the grid wires repels the electrons as they attempt to travel between the cathode and the plate. This effect prevents current from flowing. In this way, the tube with one or more grids can act as a switch. Figure 1 shows the schematic symbol for a triode with each of the components labeled. Note, that it is common to omit the heater from a diagram, and we will usually do so when describing logic circuits. 3. Simple Tube Circuits a. Tube Characteristics in Ciruits. When the grid of a tube is sufficiently negative that essentially no current flows, we call this condition cutoff. Con- versely, when the grid is sufficently positive (at 0 volts or even slightly positive with respect to the cathode), then there is no restriction on the current flow due to the grid. This condition is called saturation. When tubes are used for amplifiers, they are typically operated in the region between cutoff and saturation. However, for logic circuits, we want the tubes definitely on or off. So we operate them firmly in cutoff or saturation. b. When analyzing circuits, we take as a convention that positive current flows from a higher potential to a lower potential. This means that conventional current flows in the opposite direction to the electrons, meaning that the conventional current in a tube flows from plate to cathode. c. When a tube is operating in saturation, the current flow through the tube implies that there is a very small voltage drop between the plate and the cathode. On the other hand, when the tube is in cutoff, there can be a very large difference in voltage between the plate and cathode. We use this characteristic of the tube extensively in the logic circuits we will examine. d. Tubes as Simple Switches. In figure 2, we have two simple tube circuits where the tube operates as a simple switch. When considering these circuits, it is important to remember that Vb > Vc. We should also point out, that we will generally be assuming that the output of the circuit is fed into a high impedance circuit so that the current flowing throught the ouput is small. (1) Looking first at the non-inverting circuit (A, fig. 2), if the grid is at a negative voltage, the tube is in cutoff. No current will be flowing through the tube, and the output will be pulled down to Vc through the resistor Rc. If the grid is at a hight voltage level ≥ Vc, then the tube will be in saturation. Becuase the tube is conducting, the 2 Vb Vb Rp INPUT OUTPUT OUTPUT INPUT Rc Vc Vc A. NON-INVERTING B. INVERTING Figure 2. Triode Switching Circuits voltage difference between the plate and cathode will be small, and nearly all of the voltage dropped between Vb and Vc will be through the resistor Rc. This makes the output close to Vb. Putting it all together, when the input is low (< Vc) the output will be low (Vc) and when the input is hight (≥ Vc) the output will be high (Vb). A circuit with this configuration is often called a cathode follower. (2) The inverting circuit works in the opposite way. When the input is low, the tube is in cutoff, and no current flows. So the output is pulled up to Vb through resistor Rp. When the input is high, the tube is in saturation, and current flows to reduce the voltage drop across the tube to a small value. This makes the output approximately Vc. The inverting behavior of the circuit is summarized by seeing that a low input leads to a high output, and a high input leads to a low output. e. Grid Biasing. One problem with the circuits just presented is that they cannot be directly cascaded. That is the output of one can’t be used as the input to another. The reason is that the range of voltages that constitute high and low signals for the inputs is different than that which appears on the output. Without getting into the details, we often deal with this by also connecting the grid through a resistor to a voltage source that is negative with respect to the cathode. Section 2. Boolean Operations 4. General In modern digital design, we tend to think primarily in terms of signals that are described by expressions using the Boolean operators AND, OR, and NOT. Indeed, we often begin with the Boolean expression and then generate the circuit from that expression, building on the work of Claude Shannon in his influential master’s thesis at MIT. However, in the design of the ENIAC, we find a little different design style. Here, we find an approach based around pulses. In the following paragraphs, we see how the set of pulse manipulations in the ENIAC map to Boolean operations. 3 Vb GATE PULSE OUTPUT Vc Figure 3. Tetrode Gate (AND) Circuit 5. NOT In some parts of a circuit, we might need a pulse to be positive-going, but in other parts negative-going. In terms of the high and low logic levels, converting from one to the other is performed by the Boolean NOT. That is the primary use of the inverting circuit we saw earlier. In the contemporary documents written about the ENIAC design, the simple switching circuits are often referred to as non-inverting and inverting buffers. 6. AND As we will see in the larger-scale design of the machine, there a number of places where we wish to selectively allow or disallow a pulse to pass. In the ENIAC this is handled by a gate. Whereas today we refer to any of the Boolean logic elements as gates, the gate in the ENIAC always performs an AND function. Typically, the gate is implemented with a single tube having more than one grid. To illustrate this, consider the tetrode circuit shown in figure 3. If we apply the pulse we want to transmit on one grid and the gate signal on the other grid, we get the desired behavior. When either the pulse input or the gate input are low, the corresponding grid is negative with respect to the cathode and the flow of electrons is blocked. However, if the gate input is high when the pulse comes in, then the tube will conduct during the pulse, because both grids are high. When the output is taken from the cathode as shown here, the function of the tube is the same as the Boolean AND. On the other hand, we can get a NAND function, by taking the output from the plate.